Generative AI News - issue #4
Exploring AI’s potential and staying current with AI literacy
Hi everyone,
It was a busy month in generative AI news! Events from OpenAI, Google, and Micosoft led to lots of news stories, some of it controversial (as usual with AI news).
In this issue we’ll look at:
updates from OpenAI, Google, Microsoft, and Anthropic (which is now available in EU countries)
my 3-part series of posts in the LibTech Insights blog from ACRL: Librarians’ Guide to Answering Students’ Technical Questions about AI
and Librarians’ Guide to Answering Students’ Ethical Questions about AI. Practical questions will be published on June 3rd.
stories about AI for education and for libraries (there is so much happening)
a new Tips section, for practical tips on using generative AI
a “just for fun” section, with info about “golden gate Claude”
And much more…
Enjoy!
Foundation Model News
Open AI
ChatGPT in 4o mode doesn't have the new voice and image features yet - Simon Willison’s Newsletter
This was confusing to many people.
Using ChatGPT's Free Tier - FAQ - OpenAI
This makes it clear what you get with the free version of ChatGPT. It’s very limited compared to the paid version.
With OpenAI offering GPT-4o for free, who should be paying for ChatGPT Plus? - VentureBeat
”For instance, when using GPT-4o, free users will have certain usage restrictions. There will be a limit on the number of messages (across text and images) they can send to the model, depending on the usage and demand at that specific moment. Once the limit is reached, the chatbot will automatically switch back to GPT-3.5, reverting to the same experience.” (So there will still be a digital divide between those who can afford the paid version and those who can’t).
OpenAI didn’t copy Scarlett Johansson’s voice for ChatGPT, records show - Washington Post
But they did choose a voice actor that sounds similar to her.
‘Not an imitation’: OpenAI pauses ChatGPT voice that sounded like Scarlett Johansson - VentureBeat
You can listen to samples of the 5 voices to choose from in this 30-second video I made in March (before all of this happened). Sky is no longer available.
Google
6 Things Announced at Google I/O, One That Mattered A Lot, and One Mystery - Synthedia
The mystery is the AI eyeglasses that were shown, but not talked about.
Google takes on OpenAI’s Sora with stunning new generative AI video model Veo - VentureBeat
Like Sora, it’s not live yet. But creators can join a waitlist.
Google strikes back at OpenAI with “Project Astra” AI agent prototype - Ars Technica
The Elmer's Glue Pizza Error Is More Fascinating Than You Think - Mike Caulfield
Caulfield is the creator of the SIFT method for evaluating information. (Many librarians are familiar with his methods).
Gen AI is Way More Capable than what Google Answers suggests - Stefan Bauschard
Perplexity does better.
Microsoft
8 Announcements from Microsoft BUILD - Synthedia
They are focused mainly on enterprise uses.
Anthropic
Mapping the Mind of a Large Language Model - Anthropic
An interesting step forward in “interpretability,” or understanding what happens inside the “black box.” This research could be used to help make models safer. For a good explanation of their paper, see AI Explained on YouTube. (Start at 12 min 47 secs).
Claude is now available in Europe - Anthropic
Cohere
Cohere launches open weights AI model Aya 23 with support for nearly two dozen languages - VentureBeat
”By having access to them in an open release like this, third-party researchers can fine tune to the model to fit their individual needs.” “… serves 23 languages. This includes Arabic, Chinese (simplified & traditional), Czech, Dutch, English, French, German, Greek, Hebrew, Hindi, Indonesian, Italian, Japanese, Korean, Persian, Polish, Portuguese, Romanian, Russian, Spanish, Turkish, Ukrainian and Vietnamese.”
Generative AI for multimedia
Wirestock lets photographers and artists get paid when AI companies train on their work - VentureBeat
Hollywood at a Crossroads: “Everyone Is Using AI, But They Are Scared to Admit It” - Hollywood Reporter
The Images That Boost Support for Sustainable Transportation - Bloomberg
“AI-generated illustrations of streets with bike lanes, bus lanes and fewer cars have the power to persuade, a new study finds.”
Enabling Creativity: A Panel on AI, Art, and Copyright - Chamber of Progress on YouTube
A new note on AI Aesthetics - Lev Manovich on LinkedIn
There is no such thing as “AI aesthetics.” There are only different users who have different tastes, and different degree of visual competence.”
Is the Music Business Too Afraid of Big, Bad AI? - Variety
Suno raises $125M to become the ‘ChatGPT’ of music creation - VentureBeat
Generative AI in education
Using AI to Help Students Teach in Order to Learn - Inside Higher Ed
A.I. Program Aims to Break Barriers for Female Students - New York Times
Break Through Tech AI - MIT Schwarzman College of Computing
“Break Through Tech AI is a new program to bridge the talent gap for women and underrepresented genders in computing fields.”
Academics push for AI in service of Africa and its goals - University World News
ASU reflects on first semester with Open AI partnership - ABC 15, Arizona
New BYU computer science study shows four ways students are actually using ChatGPT - Brigham Young University News
Rethinking Plagiarism in the Era of Generative AI - Journal of Intelligent Communication
Generative AI in libraries
Generative A.I. supercharging the mission of libraries: risks and opportunities - ALA United Nations International Relations Committee Sustainable Development Goals Subcommittee
I enjoyed this free event on May 21. The keynote by Leo Lo was excellent. He emphasized that library leaders should foster a culture of experimentation, and invest in upskilling staff. There was a series of 5 minute mini-talks by library professionals from around the world. Recording available.
Imagining library futures using AI and machine learning - OCLC Research Blog
Stacking Scenarios: a layered approach to futures planning - Brian Mathews
AI in academia: Academic librarians are helping both students and instructors navigate the rapidly evolving field of artificial intelligence - Henrietta Thornton, Library Journal
I was interviewed for this article.
A new intermediary: How AI may impact libraries, research and information retrieval - Naomi DuBovis, Cronkite News
I was also interviewed for this article.
Humane Ingenuity 53: Books are Big AI's Achilles Heel - Dan Cohen, Northeastern University. See also Towards a Books Data Commons for AI Training - Open Future
Association of Research Libraries Releases Guiding Principles for Artificial Intelligence - ARL News
Libraries AI Literacy Hackathon Expands Educational Resources - Carnegie Mellons University Libraries
My offerings
The editor of LibTech Insights (blog from ACRL) asked if I’d be willing to write a series of posts with Q&A about AI. The first two are live now.
Librarians’ Guide to Answering Students’ Technical Questions about AI
Librarians’ Guide to Answering Students’ Ethical Questions about AI
A third one will be published on Monday, June 3, focusing on practical questions.
We’ve added all the FAQs in these articles to our LibAnswers system. You can find them here: Generative AI FAQs. Feel free to copy and modify them for your own library.
Experimenting with AI music generation - my blog post about Suno
I’m especially interested in AI music generation given my musical background. I have a Masters in Music from Boston Conservatory and I spend my twenties working as a performing musician in Boston.
Tips
5 prompts for making ChatGPT more accurate - Descript
Explains few-shot prompting, chain-of-thought prompting, and more.Beyond the prompt, towards AI conversations - Exponential View
How to strategically design your conversations with AI to get better outcomes - with examples.Surprising ways to prompt AI - Wonder Tools
These are fun.DALL-E 3 Full Playbook - Tianyu Xu on LinkedIn
Download the free PDF for lots of ideas for image prompting, with examples. I had fun making images of a giraffe with glasses using several of his prompt ideas. Here are my experiments.I rigged ChatGPT's new memory into my personal JARVIS - I versus AI on YouTube
A creative way to use the new memory feature of ChatGPT. And she has a fun channel: ”I versus AI”ChatGPT now lets you import files directly from Google Drive, Microsoft OneDrive - VentureBeat
For Plus, Team, and Enterprise users.How to change your language setting in ChatGPT - OpenAI
How to do this in your browser and the mobile app.How to use Google Search without AI: the ‘udm=14’ work around - VentureBeat
Gemini 1.5 Pro-powered side panel launching in Gmail, Google Docs, and more - 9 to 5 Google
Accessibility
AI & Accessibility - Cornell Center for Teaching & Innovation
Guidelines for using generative AI to create more accessible content.
‘Without these tools, I’d be lost’: how generative AI aids in accessibility - Nature
”A rush to place barriers around the use of artificial intelligence in academia could disproportionately affect those who stand to benefit most.”
AI Meets Accessibility: A Conversation with Ioana Tanase and Christina Mallon Online Class - LinkedIn Learning.
A useful mini-course.
Arizona State’s Image Accessibility Creator - Arizona State University
Upload an image, get a complete description of it, along with an alt text. I do this manually all the time with ChatGPT, so it’s nice to have this as a bookmarkable tool.
AI for OER Accessibility and Higher Education, from ChatGPT for Higher Education by Rob Rose, University of North Florida
More than alt text – create accessible tables, lists, and more.
Thought-provoking
New SRI/PEARL survey now published, reveals worldwide public opinion about AI - University of Toronto
”Globally, people feel positive towards AI. China, India, Indonesia, and Kenya display particularly positive feelings toward AI. Countries including France, the U.S., Canada, and the U.K. display the most negative views.”
Meta says AI-generated election content is not happening at a “systemic level” - MIT Technology Review
Researchers warned against using AI to peer review academic papers - Semafor
Four Singularities for Research: The rise of AI is creating both crisis and opportunity - Ethan Mollick
Can AI be used to automate social science research? - The Future of Being Human
AI as public infrastructure - Public AI Network
Using Generative AI for Human Rights Advocacy - Witness
Find more AI articles from Witness.
Positive Use Cases of Synthetic Media (aka Deepfakes) - Towards Data Science
For example: “The Dalí Museum in St. Petersburg, Florida, created an exhibition called Dalí lives, bringing him back to life using deepfakes for visitors to interact and take a selfie with surrealist painter Salvador Dalí.”
A guide to applying AI to real-world problems - Semafor
Interview with the author of the book, AI for Good.
It's Time to Update Copyright for Generative AI - Andrew Ng
Google’s AlphaFold 3 AI predicts the very building blocks of life - VentureBeat
The AI That Could Heal a Divided Internet - Time
The Ethics of Advanced AI Assistants - arxiv.org
Chapter 18 of this long paper has a nuanced view of the climate impacts, including benefits, instead of the usual fear-dominated predictions.
Internet use statistically associated with higher wellbeing, finds new global Oxford study - University of Oxford News
AI’s impact on jobs
Most in-demand skills for 2024 — hint, genAI is at the top - ComputerWorld
The New Jobs for Humans in the AI Era - Wall Street Journal
Creative AI jobs board - Curious Refuge
Mostly filmmaking with AI, mostly remote jobs.AI can restore the jobs that automation and globalization killed, a report says - Quartz
What MidJourney, DeepL, ElevenLabs, and HeyGen have in common (besides special teams and explosive growth!) - Sarah Tavel’s Newsletter
Just for fun
Golden Gate Claude - Anthropic
Read about this in the research paper mentioned at the top of this issue.
”Altering the feature had made Claude effectively obsessed with the bridge, bringing it up in answer to almost any query—even in situations where it wasn’t at all relevant.”
So for a few days, they made this weirdly adjusted model available for anyone to try. Here’s an example. You can find more examples by searching X for “golden gate Claude.”
Your own homemade GPT-4 robot, cheap? - I versus AI, YouTube
How she made a $600 DIY AI robot.
OnlyBots
A social network for bots (not humans). Learn more.
Researchers build AI-driven sarcasm detector - The Guardian
“…the group trained a neural network on text, audio and emotional content of video clips from US sitcoms including Friends and The Big Bang Theory… One scene the AI trained on was Leonard’s futile effort to escape from a locked room in The Big Bang Theory, prompting Sheldon to observe: “It’s just a privilege to watch your mind at work.”
The future
With Spatial Intelligence, AI Will Understand the Real World - Fei-Fei Li, TED - YouTube
Open Weights, Open Society, Open Future - Daniel Jeffries
Edge AI: The accessible, sustainable future of AI - VentureBeat
Using ideas from game theory to improve the reliability of language models - MIT News
Learn more
If you want to learn more about generative AI, contact me about doing a webinar for your group.
Later this year I’ll offer a 4-6 week asynchronous online course about generative AI. Sign up here if you’d like to be notified when it’s ready.
And as always, you can follow me on X, Mastodon, or Bluesky, where I post daily about generative AI.
And share this newsletter with your friends and colleagues!